Вы находитесь на странице: 1из 8

IMPACT AND MANAGEMENT OF STRAY CURRENT ON DC RAIL SYSTEMS

P. Aylott*, I. Cotton, C. A. Charalambous+

* Intertek CAPCIS, United Kingdom, pete.aylott@intertek.com University of Manchester, United Kingdom, ian.cotton@manchester.ac.uk + University of Cyprus, Cyprus, cchara@ucy.ac.cy It should be noted that the iron reduction reaction is not thermodynamically preferred and that iron does not plate back onto the rail. Corrosion of metallic objects will therefore occur from each point that current transfers from a metallic conductor, such as a reinforcement bar in concrete, to the electrolyte (i.e. the concrete). Hence stray current leakage can cause corrosion damage to both the rails and any other surrounding metallic elements. In a few extreme cases, severe structural damage has occurred as a result of stray current leakage in other cases unplanned system shutdowns for track replacement have occurred. There is therefore a stray current control requirement to minimize the impact of the stray current on the rail system, supporting infrastructure and third party infrastructure. It is significant to note that stray current has not always been perceived as a problem and has been positively encouraged. Schwalm and Scandor [1] produced a paper detailing such a view that states that rails are generally not insulated from the earth so that part of the return current travels through the earth and makes use of any metallic underground path in the vicinity that provides conductivity The essential elements of a transit system are the rails, the power supply and the vehicles. The design and placement of these elements of the transit system dictates the stray current performance in terms of the total stray current leaving the rails. If the total stray current for a given design of system is high, a stray current collection system may be needed to control the path through which this stray current returns to the substation negative bus. If a stray current collection system is not provided, considerable corrosion of the supporting infrastructure and of third party infrastructure may occur. However, as stated by Schaffer et al [2], no stray current collection system will be needed if the rail insulation and power system design themselves can keep stray current levels below a damage causing value. It is therefore obviously desirable to eliminate the need for any stray current collection system by controlling the level of stray current being produced by the transit system. Means to reduce stray current levels below a damage causing value may include measures such as increased power system cross-bonding, increased rail to earth resistances (by use of better coatings / insulating

Keywords: Rail transportation, Corrosion, Stray Current, Transit

Abstract
The production of stray currents by DC transit systems can lead to the corrosion of nearby buried metallic structures, such as rail supporting structures, pipelines and cable sheaths. This paper describes the way in which stray current is produced by a DC transit system, measures that can be taken to mitigate the production and the impact it will have on the transit and surrounding infrastructure.

1 Introduction
Current leakage from DC railway systems is an inevitable consequence of the use of the running rails as a mechanical support / guideway and as the return circuit for the traction supply current. Since the rails have a finite longitudinal, or series, resistance around 30-60 m /km or 30-60 /m of rail and a poor insulation from earth typically from 2 to 100 km a proportion of the traction current returning along them will leak to earth and flow along parallel circuits (either directly through the soil or through buried conductors) before returning onto the rail and the negative terminal of the DC rectifier. It should be noted that in a DC system, the current loss is by direct leakage. Induced effects found in AC systems are less significant in terms of corrosion damage. Given that current flow in a metallic conductor is electronic, whilst that through electrolytes such as the earth, concrete etc. is ionic, it follows that there must be an electron to ion transfer as current leaves the rails to earth. Where current leaves the rail to earth there will therefore be an oxidation, or electron producing, reaction:
Fe Fe 2 2e (1)

This reaction is visible after time as corrosion damage. For current to return onto the rail there must be a reduction or electron consuming reaction. In an oxygenated environment this will typically be:
O2 2H 2O 4e 4OH

(2)

273

supports), and the encasement of the track slab by an insulating membrane This paper describes the mechanisms by which stray current forms from DC transit systems and goes on to explain how this stray current can be minimised. Finally, the ways in which this stray current can cause damage to nearby systems is discussed.

Figure 2. Rail to earth voltage profiles for a floating and grounded rail system A positive voltage in Figure 2 represents the case where a current leaks out of the rails into the earth. For the negative voltage case, current leaks back into the rails. The magnitude of the current leaking from the rails is determined by the voltage to remote earth at any point along the track and the resistance to remote earth of each rail. At 500m down the track, the voltage to remote earth will be 0V (implying no current leakage in either direction). In the floating system, stray current will therefore leave the rails in the region 0 - 500m and then re-enter the rails in the region 500m 1000m. This is shown in Figure 3.
1000A flows into rails RUNNING RAILS 1000A collected from rails at substation

2 Stray Current Production


Figure 1 shows a 1km section of track used to illustrate the rail to earth voltage profile when a train draws current from a substation. This 1km section is representative of a symmetrical 2km section of track with a single train at the centre and a substation at each end. The 1000A that has been produced by a substation at the far-end of the track is being drawn by a train placed at 0m. For every 1m /km of track resistance, there will be a resulting voltage drop of 1V/km along the rail. Take a case where the resistance of a single rail is 40m /km (20m /km for the track). For 1000A current, the resulting voltage difference between the two ends of the track will be 20V.
1000A flows into rails RUNNING RAILS 1000A collected from rails at substation

1000m +V -V I I

Current leaves rails and enters earth Current leaves earth and re-enters rails

1000m

Figure 1. Section of model to illustrate stray current production This voltage will appear on the system in one of two ways. In a floating system where the running rails (and hence the DC negative bus) are allowed to float with respect to earth, the voltage will appear on the rails as 10V to remote earth near the train and -10V to remote earth near the substation. In a grounded rail system, where the running rails are effectively bonded to earth (via a stray current collection system or any reinforced concrete/metallic structure around the track such as a tunnel) at the substation, the voltage will appear on the rails as 20V to remote earth at the train and 0V to remote earth at the substation. Figure 2 shows the rail to earth voltage in a floating and grounded system.
20

Figure 3. Basic model of floating rail system illustrating stray current leakage In the case of a grounded rail system, where the voltage is always positive with respect to earth, stray current leaves the rails along their full length and returns to the traction system power supply at the substation earth bond (i.e. through the substation earthing system and any metallic components connected to it). For the two forms of system described, the overall stray current level can be described using the following equations. These equations are based on the single-train case shown in Figure 3 with a uniform rail insulation. In the floating system, stray current leaves the track over the first 500m returning to the track over the final 500m. It can be shown that the total stray current leaking from the system can be described as:

I stray

15

Irt l 2 8rc (3)

Rail to earth voltage (V)

10

0 0 100 200 300 400 500 600 700 800 900 1000

where I is the traction current in amps, rt is the resistance of the track (i.e. two parallel rails) in ohms per kilometre, l is the distance between the train and the substation in kilometres and rc is the resistance to earth of the tracks. For a grounded system, this equation can be rewritten as:
Distance from train (m) Floating System Bonded System

-5

-10

I stray

Irt l 2 2rc

(4)

274

The increase in stray current level by a factor of four on the grounded system arises from the doubling in the peak rail to earth voltage in combination with a halving of the resistance through which stray current can leak by a factor of two (due to doubling the amount of track at positive rail to earth potential). It would therefore seem that floating running rails are the best option if stray current is to be minimised. This conclusion is shared by Yu and Bomar. Yu [3] states that the floating rail system is the best option for the reduction of stray current levels while Bomar [4] describes a case where extremely high levels of stray current was observed in a system where the rails were directly bonded to a ground mat at the traction substations. While these results clearly demonstrate the advantages of a floating rail system, it must be proved that unsafe levels of track to earth voltages will not develop during fault conditions (such as the conductor rail coming into contact with earth). As safety is the prime concern in the design of mass transit power systems, grounded systems may occasionally be the only choice. Modern protective devices do, however, allow faults to be detected and cleared with relative ease. An oft-proposed variation on these systems is the use of a diode-bonded approach where the rail is connected to the ground mat via a diode. This diode will prevent stray currents passing directly from a ground mat to the rail. When the rail is at a negative potential with respect to earth, the system is therefore floating. The diode will, however, appear as a shortcircuit when the rail potential moves positive with respect to earth and the general effect is to increase stray current levels in comparison to a floating system [5,8]. Real systems see more complex combinations of train movements than those just described. Dynamic simulations that map train movement and rail potential are often required to describe the production of stray current across the system as a whole.

Figure 4 shows the variation in the maximum stray current leakage density of the floating system previously described in Figure 1 as a function of the resistivity of the base material and resistance of the insulating pads used to fix the track to the ground at regular intervals. The stray current leakage density is expressed in A/m, i.e. the stray current leaving a 1m section of rail. The maximum stray current leakage density is found at the location of the train or substation in the case of the floating system where the rail voltage is at a peak. In the CDEGS software [6] used to carry out this modelling, the resistance of insulating pads used in a rail system must be converted to a coating of a given resistivity that is placed uniformly along the rails. This simplifies the modeling requirements. Altering the resistivity of a 10mm thickness coating placed around the rail varies the value of insulating pad resistance. A rail coating resistivity of 100M m (the last point on the x-axis) is equivalent to an insulating pad resistance of 340 /km (produced by all the insulating pads found in 1km being placed in parallel). The figure shows that the resistivity of the material the rail is laid on does not have an effect on the stray current leakage density until the rail coating resistivity drops significantly lower than 100k m, equivalent to an insulating pad resistance of 3.4 /km. The base material resistivity would only become significant in conditions where the rail insulation was rendered ineffective (for example for light rail systems running on streets that are gritted with salt during winter). The key point illustrated by the graph is the illustration of the importance of the rail insulation. If the rail insulation resistance can be maximised then the resulting stray current leaving the running rail will be minimised.
1 Maximum Rail Stray Current Density - A/m

0.1

0.01

3 Impact Of Resistivity

Rail

Insulation

And

Soil

0.001

0.0001

An important parameter in equations (3) and (4) is the rail resistance to earth. If near perfect insulation was placed around the rails, any level of rail voltage could be tolerated with minimal stray current effects (although it should be noted that other considerations such as touch voltages restrict the maximum rail potentials allowed in a traction system). The rail resistance to earth is usually a function of the insulating pads upon which the running rails are mounted and the resistivity of the base material (e.g. concrete or ballast) on which the rails are laid. In normal circumstances the resistivity of the rail insulation / the pads upon which the rail is mounted is more significant than the resistivity of the material upon which they are placed (such as concrete).

0.00001 1000

10000

100000 Rail Coating Resistivity / Ohm-m 5 Ohm-m 50 Ohm-m

1000000

10000000

200 Ohm-m

Figure 4. Variation of maximum rail stray current density as a function of rail coating and base material resistivity In practice, the production and operation of a transit system with a high rail to earth resistance is possible in the short term after construction. Regular maintenance of the insulation is required thereafter to ensure that there is no decrease in resistance and resulting risk to the transit system or third party infrastructure. However, it may not be possible to maintain the rail to earth insulation of surface transit systems running on roadways since the insulation will quickly become coated with dirt and/or salt during winter weather.

275

4 Impact Of Stray Current On Infrastructure


As previously detailed, corrosion will occur at each point that current transfers from a metallic conductor, such as a reinforcement bar in concrete, to the electrolyte (i.e. the concrete). The previous analysis has shown that current leakage will occur from areas of the rail in both a floating system and a grounded system. The running rails are therefore the part of the rail system where corrosion must occur unless they are perfectly insulated from earth. The level of corrosion may influence the lifetime of the running rails and this should be considered at the start of any analysis. Should the running rails be placed in an area where there are no other metallic conductors, the only corrosion risk is to the rails. However, a transit system may be placed in a tunnel made with some form of reinforced concrete or for urban transit systems close to power cables, gas mains and similar utilities. For a tunnel system the primary corrosion risk is to the rails, their fixings and the tunnel walls. The exact nature of the risk depends on the form of tunnel construction (cut/cover or bored), the local soil resistivity, the volume of metallic material used in the construction and the longitudinal continuity of the construction. There is also a secondary risk to structures outside the tunnel. For an elevated system, the primary corrosion risks are to the rails and at the reinforcement and bearings at the deck / pier interfaces and where this connects to the ground at the foundations. There are then secondary risks to structures, pipes and cables outside the transit system. The exact nature of the risk depend on the reinforcement bonding at the deck interface and the manner in which the electrical safety bonding is applied. For a transit system placed at grade, the primary corrosion risk is again to the rails and their fixings as well as to any nearby utility cables/pipes. Again, the risk is determined by the local soil resistivity, the relative positioning of the transit system and the utility cables/pipes and the longitudinal conductivity of the cables/pipes. Figures 5 and 6 illustrate the risk posed to the rails and to other elements of the supporting infrastructure or third party infrastructure due to stray current flow. Figure 5 shows the effect of stray current leaving the rails, this case is found in the sections of rail systems where the rail potential is positive with respect to earth.
Running rail with insulated fastener Stray current leaves rails and enters civil / stray current control mat Civil / stray current control mat Stray current collector cable

Stray current leaves the rails causing corrosion to the rails/rail fasteners themselves. If present, the stray current will then enter any civil/stray current mat found in rail systems where the rails are placed on a concrete base. The entry of the stray current to this mat of reinforcement does not itself cause corrosion but the limited conductivity of the mat results in the further leakage of a proportion of the stray current into the surrounding soil and into any buried services (or the longitudinal conductors of any tunnel). The only corrosion risk at this time is to the mat reinforcement due to the leakage of current off the mat. Figure 6 shows the return of the stray current to the running rails. At risk in this case are the buried services and the civil / stray current mat where current flows from the metallic object into the soil. Of particular significance is the fact that in both the cases described, the civil / stray current control mat has had stray current leaving the metallic reinforcement bar to enter an electrolyte causing corrosion. The same conclusion could be applied to segments of an underground tunnel constructed with reinforced concrete.
Running rail with insulated fastener Stray current enters rails after leaving civil / stray current control mat Civil / stray current control mat Stray current collector cable

Buried Service Surrounding Soil


Stray current buried service or soil to enter civil / stray current control mat

Figure 6. Path of stray current returning to the running rails In a symmetrical system where the running rails remain floating with respect to earth, half of the system could be taken to be operating as in Figure 5 while the other half will be operating as in Figure 6. Under operating conditions this balance position will be continuously changing in response to the train movements. In a system where the running rails are grounded to earth at the substation, the stray current leakage is generally from the rails unless regeneration is taking place by a train. The stray current leaves the rails and will again pass into any civil/stray current mat and partly pass through any tunnel segments, buried services or the soil before returning to the substation earth bond via the substation earthing grid or the civil / stray current control mat (usually bonded directly to the DC negative busbar in such a case). Stray current would also not leak from the civil/stray current mat to the rails. It would, instead, pass directly through the bond between the civil/stray current mat and the DC negative busbar mentioned above. Due to the obvious corrosion risks resulting from the stray current flows detailed above, an analysis of the stray current flows within a system and the impact on the rails, the traction system supporting infrastructure and any third party infrastructure should be carried out. This assessment would usually identify the components that are vulnerable to stray current attack and calculate their lifetime based on the level of

Buried Service Surrounding Soil


Stray current leaves civil / stray current control mat and enters buried service or soil

Figure 5. Path of stray current leaving the running rails

276

corrosion that they will suffer. The calculation would involve an application of Faradays law that can relate the total charge leakage from a metallic component with the amount of metal oxidized. If the risk to any component/structure is too much too tolerate in terms of a reduction in lifetime or through safety considerations, further steps must be taken to control the level of stray current. There are a number of options such as increases in rail insulation level, an increase in the operating voltage of the traction system, the introduction of parallel conductor cables bonded to the rail and reducing the substation spacing. However, a stray current collection system is often used to 'catch' the stray current leaving the rails and provide a conductive path through which it can flow without a risk of damage to supporting/third party infrastructure.

system would operate for before corrosion started to render the structural elements of the system unsafe. Consideration of the effect of the stray current system on other buried services must also be taken in account. A CDEGS [6] model of the system described is used to illustrate the impact of factors such as soil resistivity and the size of the stray current control mat on the efficiency of the stray current collection system. CDEGS allows a geometrically accurate model of the system to be constructed and allows the investigation of the performance of a stray current collection system along its length. It is, however, restricted to the simulation of non-dynamic situations within the transit system. Figure 7 shows the perspective view of a CDEGS model. A simplified model of the stray current control mat is constructed using twelve longitudinal conductors and hoops at regular intervals. This is a reduction in the number of conductors that would be present in the real reinforcement mat. This simplification is, however, necessary since the number of conductors that would be required to model a complete reinforced concrete mat would result in excessive memory and time requirements in computations. The simplified model has the same longitudinal conductivity of the real mat and tests have shown that the simplified model performs with accuracy comparable to a more complex model. The stray current collector cable is connected directly to the stray current control mat at 100m intervals at which point the stray current control mat is also sectioned (as in practice for ease of construction). As with the dimensions of all conductors in the model, the size of the stray current collector cable can be altered to assess the impact it has on the system performance.
Running Rails Stray Current Collector Cable

5 Control Of Stray Current


A stray current collection system can be constructed under the rails in order to capture the stray current and avoid damage to other structures. Such collection systems usually take the form of reinforcement in the concrete trackbed of a traction system. This reinforcement is bonded along its length to provide a continuous and relatively low resistance path. The stray current leaking from the running rails is intended to flow into this collection system and be captured upon it as opposed to flowing through the tunnel construction or other local conductors such as utility pipes/cables. For this strategy to succeed, the mat must offer a significantly lower resistance path than segment reinforcement in a tunnel, buried services and the surrounding soil itself. In a floating system, the stray current collection system will not be bonded to the running rails. Figures 5 and 6 both show an example stray current collection system. The reinforced concrete mat placed underneath the rails is used for both structural support and as a conductive path for stray current. Connected to this mat is an insulated cable, generally copper, that increases the overall conductivity of the stray current collection circuit relative to the alternative stray current paths in the soil and other buried objects. Stray current control mats have generally been constructed in 100m sections with the starts/ends of each section being electrically connected to each other and to the stray current collector cable producing a continuous stray current path. If a designer considered that stray current was likely to be a problem in a specific region of the transit system, local stray current collection systems could be used but careful attention would have to be given to the design of the terminations of the system where severe corrosion could occur. Prediction of the stray current collection system efficiency before construction of a transit system is essential to ensure that the stray current levels will not have an adverse effect on the transit system lifetime. The restriction on the transit system lifetime is broadly based on the amount of time the

Civil/ Stray Current Mat

Figure 7. CDEGS Model Of Stray Current Collection System The running rails are placed above the mat, the separation of the reinforcement bar and the rails being equivalent to that in a real system. The running rails are simulated by a cylindrical conductor having the same longitudinal resistance as an actual rail. The rails are also coated with a resistive coating to model the insulated pads on which the rails are placed in a real system. The civil/stray current control mat is modelled as being placed within wet (30 m) or dry (180 m) concrete while a varying soil resistivity can then in turn be placed around this concrete. Buried services, tunnel segments and other

277

conductive objects that may need to be considered in any particular case can also be simulated if necessary.

diameter of the reinforcement bar used for the stray current control mat. 5.2 Variation Of Efficiency With Mat/Cable Size

5.1 Variation Of Efficiency With Soil Resistivity A model of a 1km section of floating rail system is initially used to illustrate the effect of soil resistivity on the performance of a stray current collection system. The copper stray current collector cable has a cross-sectional area of 120mm2, the steel stray current mat has a cross-sectional area of 1600mm2 and the concrete resistivity is 180 m. The soil resistivity is 100 m. The rail voltage profile due to the injection of 1000A into the track is as shown in Figure 2. The total stray current leaving the rails in the first 500m section is 26mA (equivalent to a track to earth resistance of 96 km ). This 26mA, 13mA from each of the running rails, will flow into the stray current collection system. It will then either remain in the stray current collection system or will flow into the soil surrounding the reinforced concrete slab. Measurements of the current flows are taken at the 500m point where, for this symmetrical floating rail system, no stray current is entering or leaving the rails. A plot of the efficiency (i.e. the percentage of stray current found on the stray current collection system at 500m compared to the total stray current) of the stray current collection system against the resistivity of the soil surrounding the reinforced concrete slab is shown in Figure 8. As the soil resistivity increases, the percentage of the stray current retained on the stray current collection system is increased. This is due to the reduction in the conductivity of the alternative path through the soil. The ratio of the current carried by the stray current mat in comparison to the collector cable is approximately 1:1 (not shown on graph).
100.00

In the model described, the total cross-sectional area of the steel within the mat is 1600mm2. Adjusting this crosssectional area to take into account the fact that the steel used has a resistivity 13.1 times that of copper, this would equate to 123mm2 of copper. The simulations carried out for Figure 9 show the relationship of the stray current collection system efficiency to the stray current collector cable cross-sectional area. This plot relates to a surrounding soil resistivity of 10 m where the overall efficiency of the system is relatively low (69% for a 120mm2 collection cable as compared to 90% in 100 m soil). A 120mm2 collector cable results in an approximately equal current flow in both the mat and the collector cable. This confirms that the 1600mm2 of steel within the mat is equivalent to approximately 120mm2 of copper. As the crosssectional area of the stray current collector cable is increased, the percentage of the total stray current flowing through the collector cable also increases but the current flowing through the stray current mat is decreased.
80.00

Percentage Of System Stray Current Flowing In Stated Object

70.00

60.00

50.00

40.00

30.00

20.00

10.00

0.00 120 180 240 300 360 420 480 540 600 Stray Current Collector Cable CSA / square mm Overall Stray Current System Stray Current Collector Cable Stray Current Mat

Percentage Of System Stray Current Flowing In Stated Object

90.00

80.00

Figure 9. Efficiency of the stray current collection system against stray current collector cable cross-sectional area (10 m soil around the dry concrete base). In the case where the stray current collector cable size changes from 120mm2 to 240mm2, there is an increase in the conductivity of the total stray current collection circuit by 50%. The efficiency of the stray current collection system is, however, only increased by a relatively small 3.9%. It can therefore be concluded that control of stray current levels using stray current collection systems can therefore be difficult in areas where there is a low soil resistivity or other highly conductive infrastructure.

70.00

60.00

50.00

40.00

30.00

20.00 1 10 100 Soil Resistivity / Ohm-m 1000 10000

Figure 8. Efficiency of the stray current collection system against soil resistivity While the resistivity of the soil is an important factor in the determination of the stray current collection system efficiency, it may in fact vary throughout the year and a worst case value should be taken. Factors that can be controlled are the stray current collector cable cross-sectional area and the

6 Corrosion Impact of Stray Current


Stray currents will enter third party structures and be conducted along it for some distance before leaving to return to the traction substations. It is at this point of departure from the third party structure that corrosion will occur. The

278

magnitude of corrosion damage is directly related to the current density (that is current per unit area) at the point of departure and the time over which the current flows; as opposed to the absolute current leaving the structure. The higher the current density the greater the resultant corrosion damage to the structure. For structures such as pipes and cables running parallel to the railway the locations of entry and departure can be distributed along the structure; for pipes crossing the railway at discrete locations the stray current entry and exit locations can be localised around the crossing point. It is therefore possible to sustain significant corrosion damage, even at relatively low values of stray current, because the point at which this current leaves a third party structure is often of small surface area with the resultant current density relatively high. If the magnitude of stray current is known it is possible to estimate the resultant corrosion damage over a period using Faradays electrolysis formula. Some examples of damage are given in Table 1. Metal Loss g/year cm3/year 1000 9100 1152 500 4450 576 50 455 58 Table 1 Amount of stray current versus metal loss Corrosion rates or corrosion damage are commonly assessed in terms of a corrosion rate expressed in terms of section loss per year. At a more fundamental level corrosion rates are expressed in terms of the current flow per unit area or current density, with the current flowing from the metallic component into the surrounding environment (water, soil or concrete etc) as it is the quantity of current flow that is a direct measure of corrosion rate. Current density data can be converted to section, or indeed weight, loss using Faradays law. In some cases the impact of corrosion needs to take account not only of simple section loss to the metallic component but also the consequences for other components of the resulting corrosion product. This is particularly the case with reinforcement in concrete where corrosion products may have a much higher volume than the original metal. When surrounded by concrete these corrosion products are confined and the resulting expansive forces may cause damage to the concrete which has poor tensile strength. Such damage may include cracking, spalling and delamination of the concrete and this damage can manifest itself with relatively little section loss of the reinforcement, thus even low rates of corrosion can be detrimental. The risk of such secondary damage is also dependent on the type of corrosion product formed and, to a considerable extent, this is dependent on the environment to which the metal is exposed. In the case of the subway tunnels and surrounding environment there are two basic environments that need to be considered: Amount of stray current (mA)

Environments that are essentially oxygen depleted in which non-expansive corrosion products would be expected to form. These are less likely to cause damage to the surrounding concrete although section loss of the reinforcement will still occur. This type of environment will prevail in underground conditions for example outside tunnels and may occur at the base of surface trackbeds, or in areas where concrete is saturated with water. Environments where there is a ready supply of oxygen. In these conditions the corrosion products formed are more likely to generate expansive forces, if confined, and there is a greater risk of concrete degradation as well as section loss. Such environments are likely to occur, for example, within tunnels where concrete is unsaturated with water. Ordinarily steel reinforcement in concrete is maintained in a passive condition because of the high pH of the concrete, under these conditions corrosion rates are negligible and the risk of damage to either the reinforcement or concrete can be ignored. However, when this passivity is locally lost for any reason (such as penetration of chloride ions to the bar or stray currents leaving the bar) corrosion may occur. Where concrete is confined and a ready supply of oxygen is available the section loss required to cause cracking of the concrete is low, typically in the range 150 to 300 microns. For these conditions the limiting annual corrosion rate for a 100 year design life would be of the order of 1.5 microns/year. For steel that is not placed in concrete the tolerable damage is dependent on how much section may be lost, either locally or globally, prior to the loss of structural integrity. This is difficult to determine in general terms as the tolerable section loss is to, varying, degrees dependent on the location and function of the component and the type of corrosion damage that may occur. For example a heavy structural section will be more tolerable to both localised pitting corrosion and general uniform corrosion compared to a thin walled pressure pipe where even a small amount of local pitting could have catastrophic consequences. If we take track-bed reinforcement as an example, a calculation of metal loss arising from stray current leakage is determined from Faradays laws, which for steel gives a relationship that 1 amp.year of current will corrode approximately 9.1 kg. Applying this relationship to a 12 mm rebar in oxygenated concrete, a continuous stray current density of 1286 A/m2 will be sufficient to cause cracking over 100 years under worst case conditions. This assumes that current is lost from 50% of the bar circumference (i.e. the side of the bar closest to the concrete surface). While this value seems reasonably low, if this figure was applied to a trackbed containing 25 reinforcement bars, the track would need to produce stray current at an average rate of 1.2mA/m of length to cause this level of damage. With the use of good rail insulation and stray current control systems that are carefully monitored, maintaining stray current below this level should be achievable.

279

7 Conclusions
The total stray current produced by a DC mass transit system will adversely affect the transit system itself and third part infrastructure by causing corrosion. The stray current produced by a DC mass transit system can be limited in a number of ways including the reduction of substation separation, rail resistance and operating currents and an increase in the rail resistance to earth. A system where the running rails float with respect to earth produces roughly four times less stray current in comparison to an equivalent grounded system when static simulations are carried out. Dynamic simulations show that a floating system can actually result in even higher local reductions in stray current level in comparison to a grounded system. Reducing stray current levels is best done by careful control of factors such as substation spacing, rail to earth resistance and rail resistance. However, if after analysis the stray current level produced by a transit system is too high and may affect supporting or third party infrastructure, a stray current collection system may have to be considered. The role of the stray current collection system is to collect the stray current leaving the rails and to conduct it along the traction system to the point where it re-enters the running rails. In this way, corrosion damage to supporting and third party infrastructure might be avoided. The performance of a stray current collection system is however, highly dependant on the conductivity of the system itself and of the neighbouring soil. Extremely high efficiencies can be achieved when the material surrounding the stray current collection system is highly resistive. At lower soil resistivities, the results showed the difficulty in achieving a stray current collection system with a high efficiency. In such cases it may be more economic to consider the other ways to reduce the stray current level at source (i.e. from the rails) as previously described.

[5] Case, S. DC Traction Stray Current Control - Offer a Stray a Good Ohm?, Proceedings of IEE Seminar, 21 Oct. 1999 [6] CDEGS Software, Safe Engineering Services & Technologies Ltd. [7] Y.Cai, M.R Irving, S.H Case Iterative Techniques for the solution of complex DC-rail-traction systems including regenerative braking, IEE Proceedings. Generation, Transmission, Distribution, Vol. 142 No.5, September 1995 [8] J.G. Yu, C.J. Goodman Stray Current Design Parameters for DC Railways, Railroad Conference, 1992. Proceedings of the ASME/IEEE Spring Joint, March 31 - April 2, 1992

References
[1] Schwalm L H, Sandor J G. Stray Current The Major Cause Of Underground Plant Corrosion, Materials Performance Vol. 6, 1969 [2] Shaffer R E, Smith A V, Fitzgerald J H III. Stray Earth Current Control Washington, DC Metro System, Materials Performance April 1981 [3] Yu J G. The Effects Of Earthing Strategies On Rail Potential And Stray Currents In DC Transit Railways, Proceedings of The International Conference On Developments In Mass Transit Systems, 20th 23rd April 1998 [4] Bomar H E, Dean R O, Hanck J A, Orton M D, Todd P L. Bay Area Rapid Transit System (BART), Materials Performance, December, 1974

280